Utilizing image processing techniques for detecting breast abnormalities in thermography images

Abstract

Thermal Infrared (TIR) imaging of breasts involves a non-invasive, non-ionized, passive, safe and painless scan of the breasts. It is a graphing of the changes in breasts skin temperature using thermography. Thermograms are temperature distribution patterns with differ-ent colors to indicate temperature of the different regions within the tested breast, each color refers to a certain temperature range. In this work, three breast thermography images: one for normal case and two for cancerous cases, were employed to test the performance of the proposed segmentation methods: Region growing; clustering (K-means and FCM) algorithms and Histogram based enhancement tech-nique to segment, detect and isolate the suspicious abnormal regions. These techniques were performed with the aid of suitable morpho-logical operations to get the refined regions of interest. The results proved the efficiency of the proposed techniques to extract the abnor-mal (of high temperature) regions.